Procedure to automatically interpret trial results aaa 

A random forest (decision tree) relating Optical Density readings to
binary yeast growth (growth or no growth) constructed using the R
Statistical computing language was used to determine the growth
outcomes for experiment plates in the AAA rerun. The random forest was
generated from a set of training data comprising experiment plates
from previous ADAM experiments. A growth outcome is found for each
well on the plate and the majority growth outcome is given to a set of
wells belonging to the same trial test. The growth outcome
for the set of wells constituting the knockout strain and added
nutrient will be used as a training example for the hypothesis
generation step.